Challenge of AI-Based Ophthalmic Screening in Underserved Markets

Artificial intelligence has transformed numerous industries, yet its integration into healthcare remains a complex challenge, especially in markets with structural inefficiencies. PROSPERiA, a Mexican startup, developed retinIA, an AI-powered software designed initially to detect lesions associated with diabetic retinopathy (DR). However, introducing this technology into Mexico, a country with one of the highest diabetes prevalence rates but limited preventive healthcare awareness, required more than just an innovative product. It demanded an in-depth understanding of the local healthcare ecosystem, adaptation of the business model, and engagement with multiple stakeholders.

The Context: Diabetes and Vision Loss in Mexico

Mexico is among the countries with the highest incidence of diabetes, a disease affecting between 15.3% of adults (International Diabetes Federation, 2023). Additionally, over 10 million Mexicans are estimated to be living with prediabetes, many of whom will develop the disease without timely intervention.

One of the most severe yet underdiagnosed complications of diabetes is DR, a condition that causes microvascular damage to the retina due to chronic high glucose levels and can lead to blindness. Studies indicate that up to 35% of people with diabetes will develop some form of DR, with 7% progressing to vision-threatening stages. Despite this, less than 30% of diabetic patients in Mexico undergo annual ophthalmic screening.

The reasons behind this lack of screening are multifaceted: low awareness at the primary healthcare level, where people with diabetes have more frequent contact; cost barriers for those without public health coverage; and a lack of preventive referrals within the public health system to specialized hospitals. Additionally, the entire healthcare system prioritizes acute care over prevention, which poses a serious challenge for diseases like diabetic retinopathy. Patients often do not experience visual symptoms until retinal damage is already advanced — and, in most cases, irreversible. These gaps create an urgent need for scalable, cost-effective, and accessible screening solutions — precisely what AI-powered tools can provide.

Challenge No. 1: Bridging the Preventive Health Education Gap

The biggest challenge was not just introducing an AI solution but also educating people about retinopathies and changing mindsets regarding the importance of early detection. Many patients and general practitioners underestimate the need for ophthalmic screening until symptoms become severe. This results in a reactive approach to eye health rather than a proactive one.

Our strategy to tackle this challenge was to work with a public relations agency to enhance media coverage about our company, our innovative solution, and the problem we address. On the commercial side, we targeted the most common points of contact for people living with diabetes. We engaged with the public health system and medical offices adjacent to pharmacies, which provide low-cost medical attention where people can easily seek care for symptoms and purchase medications.

This strategy led us to two main learnings:

• Relying solely on our entry into the public healthcare system would take a very long time and a lot of resources, both of which are often not available for startups at an early stage. Therefore, in parallel with our public health and government efforts, we needed to identify “quick-close” potential clients, such as small and medium-sized clinics.

• We needed to rebuild a strategy focused on a stronger willingness to pay. Who would be most interested in early treatment for eye-threatening diseases?

Challenge No. 2: Multi-User Complexity and Adoption Resistance

Unlike many healthcare software solutions designed for a single end-user, retinIA had multiple users, each with different concerns:

• Patients need clear and digestible information in their results report. The report must clearly indicate whether they should see a specialist, which type of specialist, and how urgently.

• Operators (mainly opticians or nurses) require a user-friendly, seamless platform. In addition to being easy to use, the platform must include tools that help them guide patients through the screening process and explain the results effectively.

• Physicians (general practitioners, ophthalmologists, endocrinologists) need to understand both the result and the rationale behind it. If the results lack explainability, physicians may struggle to trust and adopt AI-aided screening.

To apply the principles of human-centered design, all users must be considered, their needs identified, and solutions tailored to meet them as effectively as possible. In addition to continuously improving the platform based on user feedback, we developed training videos, frequently asked questions on our webpage, disease information linked to the report, and explainability tools, including enhanced retina images and heat maps highlighting potential lesions.

Challenge No. 3: From SaaS to Health Campaign Implementation

Initially, our vision for retinIA was to operate as a pure software-as-a-service (SaaS) company, providing AI-driven retinal screening to healthcare providers. However, we quickly realized that this model posed significant adoption barriers. Many potential clients hesitated to invest in retinal cameras or train personnel to conduct screenings, limiting the reach and impact of our solution.

To overcome this, we pivoted from solely offering software to directly implementing vision health campaigns. Instead of relying on external adoption, we took a hands-on approach, hiring and training our own opticians to operate the cameras, use the platform, and provide patient guidance. By executing these health campaigns on behalf of our clients, we eliminated the main barriers that prevented them from adopting our technology: the cost of equipment and the need for trained personnel.

This shift allowed us to work with a broader range of organizations, including companies, clinics, associations and public health initiatives that wanted to offer vision screening but lacked the infrastructure. Through this model, we not only ensured the effective use of our AI solution but also expanded access to preventive eye care, directly addressing the gaps in ophthalmic screening in our target markets.

Lessons Learned and Future Opportunities

Our experience with retinIA provided valuable insights applicable to any AI-driven healthcare solution:

• Know your ecosystem: Technology alone does not drive change — understanding users and stakeholder dynamics does.

• Educate, don’t just sell: Adoption increases when users understand why the technology matters.

• Be flexible with business models: A single approach rarely works; diversifying channels accelerates adoption.

• Validate and iterate: Early pilots and continuous improvements are essential for sustained success.

Introducing AI in healthcare is not merely a technical challenge, it’s a behavioral, economic, and systemic one. Our experience underscores the importance of combining technological innovation with deep market understanding and business agility. As AI continues to revolutionize medicine, startups like ours must remain adaptive, patient-focused, and committed to making healthcare more accessible and effective for all.